[PLASKETT-DELURY] DETERMINATION OF THE SOLAR ROTATION 119 
published. There are four essential precautions necessary in obtaining 
accurate values of the solar rotation. 
1. The emulsion on the photographic plate must be exactly in the 
focus of the spectrum. 
2. The illumination of the grating from the opposite limbs of the 
sun must be similar and uniform. 
3. The solar definition must be good, and the sky free from haze. 
4. Care must be taken that the reflecting prisms receive light from 
the latitudes desired. 
1 and 2 above, conditions inside the spectrograph, are essential in 
preventing spurious displacements of the spectrum lines producing 
systematic errors in the rotation values. If either condition is exactly 
fulfilled the other is not so essential, but as it is impossible either to 
keep the plate at the exact focus or to have absolutely equal and uniform 
illumination from the two limbs, the only safe procedure is to fulfill 
both conditions as closely as possible. The focus may be determined 
either by the definition test or preferably by the Hartmann method of 
extra-focal exposures. 
It has been found here that in the region around À 5600 the focal 
surface is neither exactly normal to the axis nor is it a plane. The 
normal to the surface is inclined about 1° to the normal to the axis, and 
the focus is about 2-5 mm. longer at the centre than at the ends of a 
plate 30 cm. long. Consequently the plates are bent to the proper 
curvature and inclined at the proper angle, so that the whole length of 
spectrum may be in focus. The equality of illumination is tested by 
opening the slit sufficiently wide to see, through the door at the rear of 
the spectrograph, the circle of light thrown on the collimator by the 
concave mirror. In this way the prisms can be so adjusted that the 
two images are superposed and uniform illumination much more closely 
obtained than when one observes with a slit of normal width the grating 
surface from the position of the plate. 
3 and 4 are essential conditions outside the spectrograph. The 
former is maintained by exposing the mirrors to the sun only during 
the actual exposures, and shading them for a sufficiently long interval 
between exposures to allow them to return to their normal state. The 
latter is arranged by careful determination, first, of the exact position 
on the outer plate at which the light is incident, which passes through 
the slit from each limb, and second, that this position is exactly placed 
on the desired latitude. The position angle is obtained, by the aid of 
tables of the position angle of the sun’s axis, from the East and West line, 
determined by stopping the coelostat clock and allowing the sun’s 
image to drift across the face of the spectrograph, as mentioned above. 
